Re: Finding and Reminding, tech issues, 3.0 and beyond

On 15/04/10 15:05, Bastien Nocera wrote:
On Thu, 2010-04-15 at 13:05 +0100, Martyn Russell wrote:
It's neither, and that means that you need to learn a new query

I couldn't this time last year either (and that's not a benchmark for
how long it takes to learn either).

I don't think your case is a good one, given that you worked on Tracker
almost exclusively during that time.

I think Owen's point was quite pertinent here and I did say my time with SPARQL is not be a benchmark for how long it takes to learn.

Well, other than the online docs we are constantly improving:

Huh. Go to that page, and find me the ontology that pertains to media,
or contacts.

I never said it was perfect. :)

This is generated and the documentation is a work in progress. I agree, looking at it for the first time is a bit cryptic and that needs to be addressed. But generally, if you have read about Nepomuk before our online documentation, you would know how these things relate.

Our online documentation is really there to to show where we extend or differ from the original specs. The issue here is, we don't implement everything (because it isn't all useful or efficient) and we do extend some ontologies.

We will try to focus on improving this area.

The "big picture" is actually unreadable on my big
(1680x1050) screen.

That's not big :P, looks fine here on my 2560x1600 30" screen ;)

That picture is again generated and not necessarily that useful. It just shows how things are connected.

Plus our code base is riddled with examples which have to work.

In addition to all this, we are on IRC to help too.

Bastien, if you think something else is missing, please tell us and we
will try to fix it.

Am I allowed to extend ontologies?


What's the process for that?

You change data/ontologies/* and make install (or add your own to the install location).

Philip has been doing a lot of the work here, 0.9. has much better support for updating/migrating.

happens if I want to store private data (something not covered by the
existing ontology, say, the ASIN or MusicBrainz ID for music)?

You have to store it yourself. Some data doesn't pertain to all applications. That isn't to say we won't add such a feature, but right now, we don't have such a thing.

What pitfalls should I look for when I create queries?

This is quite a good thing to document. Right now we help people on a case by case basis. But generally using OPTIONAL (where a outer joins are required at the SQL level) can cause performance issues.

How do I deal with escaping in those queries?

libtracker-client has APIs for that.

Where do I find information about the "big picture" of how all those
things work together.

You mean like an end-to-end illustration for how the user starts a query and what tracker does with it to the result?

Or something else?

Apple has a nice introduction that could be used as an example of what I
would be expecting:

I will take a look, thanks.


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]